[arin-ppml] ARIN-prop-157 Section 8.3 Simplification

Michael Sinatra michael+ppml at burnttofu.net
Wed Sep 21 19:58:16 EDT 2011

On 09/21/11 16:48, David Farmer wrote:
> I think you may have a valid argument for not allowing specified IPv6
> transfers there, and even a stronger argument for not allowing inter-RIR
> specified IPv6 transfers. However, is there an equally strong argument
> for not allowing ASN transfers, especially 2-Byate ASNs?

No.  Well, there may be a strong argument, but I am not making it here.

> While we do have 4-Byte ASNs now and they are more or less compatible
> with 2-byte ASNs, much more compatible than IPv6 is with IPv4. I'm not
> completely sure the transition to 4-Byte ASNs is actually going any more
> smoothly than the transition from IPv4 to IPv6 and there may be similar
> argument to allow the transfer to 2-Byte ASNs as for IPv4.

The scope of the transition is smaller, so I would suspect that it would 
go smoother, but I don't see reason to preclude 2-byte ASN transfers 
where they are needed and justified.  I can imagine scenarios where a 
4-byte ASN might be traded for a 2-byte ASN.  I don't see a good reason 
to restrict that.

That brings me to the answer to John's question: I am opposed to the 
inclusion of IPv6 in this proposal, and am therefore opposed the the 
proposal as written.  If IPv6 is removed from the scope of this 
proposal, I would probably support it.


> I'd be interested opinions regarding that issue too.
> On 9/21/11 17:40 CDT, Michael Sinatra wrote:
>> In addition, by allowing IPv6 space to transfer in this manner, the
>> careful and sparse allocation methods ARIN and the other RIRs have been
>> doing in order to maximize some semblance of aggregation will become
>> less effective. A good chunk of the fragmentation in IPv4 space is due
>> to address blocks that have been acquired over the course of time
>> (sometimes through M&A) that are no longer aggregable. Keeping ARIN in
>> control of IPv6 space can help ensure (or at least it won't undermine)
>> the goal of having a minimum number of aggregable IPv6 prefixes
>> announced per ASN. We know that transfer policies will increase
>> fragmentation in IPv4 but we think they're a necessary evil. They're not
>> a necessary evil in IPv6.
>> michael
>> On 09/21/11 14:56, Owen DeLong wrote:
>>> I disagree. While we did not feel it was appropriate to limit IPv4
>>> transfers to
>>> legacy space, the existence of legacy space really is the only reason
>>> we needed
>>> an 8.3 transfer policy. Space covered by RSA should be returned to
>>> ARIN and
>>> the recipient should get their space directly from ARIN without the
>>> need for
>>> directed transfers.
>>> Owen
>>> On Sep 21, 2011, at 9:01 AM, Matthew Kaufman wrote:
>>>> Support. There is no reason for IPv4 to be special here. As IPv6
>>>> becomes more prevalent, we will undoubtedly see cases where someone
>>>> wants to transfer a block of IPv4 space *and* the associated IPv6
>>>> space without selling a portion of their business along with it.
>>>> Matthew Kaufman

More information about the ARIN-PPML mailing list